Developed electrophysiological algorithms—specifically online spike sorting—in C++/CUDA for researchers at the Moore Lab, Stanford University, focusing on low-level performance optimizations to enable real-time processing of high-volume neural data. Adapted Kilosort4 for online operation—processing only local recording information to support streamed data—with implementations in both C++/CUDA and Python.

Implemented additional research-pipeline components, including inter-device networking and machine-learning techniques for decoding visual stimuli from neural activity.

Built an experimental pipeline integrating Dorna Robotics hardware: leverage Dorna’s API to orchestrate precise, time-sensitive robotic-arm movements and combine computer-vision methods for autonomous object manipulation.

Analyzed and provided user interactions for large language models (LLMs) on various software engineering topics to improve the effectiveness of end-user interactions with AI.

Privately tutored upwards of 10 students in various math and computer science topics, including introductory C++/Python/Java, cryptography, and generative AI.

Planned and held weekly discussion sections, which involves preparing material for a 2-hour lecture on the week’s topic and practice problems for upwards of 100 students.

Led tasks such as homework and test grading, midterm and final review sessions, and office hours twice a week.